Improved Matrix Multiplication by Changing Loop Order

Table 6

Parallel code matrix-matrix multiplication in C/C++.

N	C++ (serial code) [i, k, j]	C/C++ (parallel code) [i, k, j]
500	0.4378	0.093
1000	2.815	0.456
1500	20.814	2.845
2000	40.748	8.102
2500	115.456	16.324